Models and issues in history-dependent mainshock hazard

摘要: The purpose of this paper is to critically review some history-dependent probabilistic mainshock occurrence models with respect to probabilistic seismic hazard analysis (PSHA). In PSHA, earthquake occurrence is usually referred to homogeneous Poisson process (HPP), which is memory-less; however, when a single fault is of concern and/or the time scale is different from that of the long term, historydependent models, may be appropriate. In the study, two categories of models are reviewed:(i) renewal processes, which are used to model the sequence of interarrival times of characteristic earthquakes, and (ii) simple Markov renewal processes, in which beyond earthquake occurrence, modeling of correlation between magnitude and interarrival time is also explicitly addressed. As an illustrative application, the Paganica fault (in central Italy) is considered to compute the seismic hazard, in terms of exceedance probability of a ground motion intensity measure value, according to each of the models rendered comparable by means of working hypotheses. Results show that as larger is the time since the last earthquake the more different hazard is implied by the different models, rendering critical their selection for sources where the time since the last event is large.